What to expect when translating into English
Translating into English usually shortens the text compared with Romance, Slavic or Germanic source languages, so the translated voice rarely has to speed up to fit the original timing. Names, product terms and numbers carry over unchanged.
From Spanish to English
Spanish to English typically shortens the text, so the English voice fits easily inside the original timing. All major Spanish accents transcribe well, which gives the translation a strong starting point.
This direction is common for researchers, journalists and companies reviewing Spanish-language interviews, focus groups and customer calls. If you only need to understand the content, a translated English transcript at 1 credit per minute is often all you need.
Regional slang and fast colloquial speech can produce literal translations. For quotes you'll publish, compare the English line with the Spanish transcript, which is delivered alongside it.
A US journalist translates a 33-minute Spanish interview recorded in Mexico City into an English transcript and checks quotes against the Spanish original.
Cost: 1,650 credits ($1.65) with a English voice, or 33 credits (3.3¢) for English subtitles only. Spanish is detected automatically; you only choose English.
From Hindi to English
Hindi to English handles Hinglish well: English words already in the speech pass straight through. Formal Hindi lectures transcribe cleanly and translate into readable English.
Indian creators and educators use this direction to reach audiences outside India and the diaspora who prefer English. Because Hindi speech often already mixes in English terms, those words come through naturally in the English output.
Bollywood songs, background music and crowd noise reduce accuracy — for lectures and talks recorded with a microphone, results are strongest.
A Delhi coaching institute translates a 22-minute Hindi physics lecture into English for students in other states.
Cost: 1,100 credits ($1.10) with a English voice, or 22 credits (2.2¢) for English subtitles only. Hindi is detected automatically; you only choose English.
From Japanese to English
Japanese to English has to add subjects that Japanese leaves out, so the translation sometimes guesses 'he', 'she' or 'we' wrongly. English text is longer than the Japanese character count suggests.
Japanese often omits subjects and objects that English needs, so the translation has to infer them. In conversations this can produce the wrong pronoun; in narrated content where context is clear, it's rarely a problem. Review lines where 'he', 'she' or 'they' matter.
Japanese creators — cooking, crafts, gaming, travel — translate into English to reach the large international audience for Japanese content. Keep Japanese terms your audience knows (like dish names) consistent in the English transcript.
A Kyoto ceramics studio translates an 18-minute Japanese workshop video into English for overseas students.
Cost: 900 credits (90¢) with a English voice, or 18 credits (1.8¢) for English subtitles only. Japanese is detected automatically; you only choose English.
From Chinese to English
Mandarin to English expands the text — English needs more syllables for the same meaning — so the dubbed voice may speed up in dense passages. Homophone errors in the Chinese transcript can produce odd English lines; check names.
Speech recognition is tuned for Mandarin, so Mandarin recordings give the strongest results; Cantonese and other varieties are less reliable. Chinese has many homophones, so names and specialised terms in the Chinese transcript are worth checking before trusting the English.
Businesses use this direction to review supplier calls, training material and market research from China; academics use it for interviews and lectures.
A sourcing team translates a 27-minute Mandarin factory-tour video into English to brief colleagues who don't speak Chinese.
Cost: 1,350 credits ($1.35) with a English voice, or 27 credits (2.7¢) for English subtitles only. Chinese is detected automatically; you only choose English.
From Korean to English
Korean to English reorders sentences and drops speech levels, which have no direct English equivalent. Honorific nuance is lost — usually fine for business and education.
Korean content — education, beauty, gaming and entertainment — has a huge international audience. Korean speech levels and honorifics don't map onto English, so the translation conveys meaning but flattens formality.
Korean puts the verb at the end, so long sentences are fully restructured. Subtitle cues follow the Korean speech timing, which can make some English cues feel slightly ahead of or behind the speaker's emphasis.
A Seoul language academy translates a 14-minute Korean study-tips video into English for international learners.
Cost: 700 credits (70¢) with a English voice, or 14 credits (1.4¢) for English subtitles only. Korean is detected automatically; you only choose English.
From French to English
French to English typically shortens text, so the English voice fits comfortably. Idioms are sometimes translated literally — check expressions in marketing content.
French and English share many cognates, which helps machine translation — but faux amis (actuellement means 'currently', not 'actually') occasionally slip through. A quick read of the English transcript catches them.
African, Canadian, Belgian and Swiss French accents transcribe well. Fast colloquial speech with verlan or slang is harder.
Numbers deserve a check: French counts some tens differently (soixante-dix is 70, quatre-vingt-dix is 90, while Belgium and Switzerland say septante and nonante), and decimal commas become decimal points in English. Prices and statistics are where a misheard number does the most damage.
A Paris museum translates a 20-minute French curator talk into English for international visitors on its website.
Cost: 1,000 credits ($1.00) with a English voice, or 20 credits (2¢) for English subtitles only. French is detected automatically; you only choose English.
From German to English
German to English shortens sentences and unpacks compound nouns. Because German verbs often come last, long German sentences can split awkwardly across subtitle cues.
German's long compound nouns are unpacked into several English words, and verb-final clauses are reordered. The English is usually shorter, so the voice track fits easily.
Engineering, manufacturing and automotive companies use this direction for training and product videos. Technical terms are generally translated well, but check internal jargon and part names.
German formality (Sie) disappears in English, which has a single 'you', so the English often reads warmer than the original. German speakers also say numbers units-first (einundzwanzig is 'one-and-twenty'), which occasionally trips up figures in fast speech — double-check numbers in the German transcript.
A Munich engineering firm translates a 24-minute German maintenance training into English for its US plant.
Cost: 1,200 credits ($1.20) with a English voice, or 24 credits (2.4¢) for English subtitles only. German is detected automatically; you only choose English.
From Arabic to English
Arabic to English works best when the speaker uses Modern Standard Arabic; dialect-heavy speech is transcribed less reliably, and transcript errors carry into the translation.
Formal speech — news, lectures, sermons, official statements — is close to Modern Standard Arabic and transcribes most reliably. Everyday dialect, especially Maghrebi, is harder, and errors in the Arabic transcript carry over into the English.
Researchers, NGOs and journalists use this direction for interviews and field recordings. For published quotes, have an Arabic speaker verify the key lines.
An NGO translates a 31-minute Arabic community meeting recorded in Amman into an English transcript for its donors' report.
Cost: 1,550 credits ($1.55) with a English voice, or 31 credits (3.1¢) for English subtitles only. Arabic is detected automatically; you only choose English.
From Portuguese to English
Portuguese to English shortens text. Brazilian and European Portuguese both work as sources; fast European Portuguese may need more transcript corrections.
Brazilian and European Portuguese both transcribe well. Brazilian creators use this direction to reach English-speaking audiences, and companies use it to review Brazilian customer interviews and sales calls.
Portuguese typically runs longer than English, so the English voice fits comfortably inside the original timing.
Watch for false friends that machine translation sometimes passes through: 'pretender' means to intend, 'puxar' means to pull, and 'atualmente' means currently. Brazilian business speech also borrows many English words ('deadline', 'feedback'), which carry through unchanged.
Portuguese uses decimal commas and dots for thousands (1.500,50), the opposite of English. The translation normally converts these, but check figures in financial content against the Portuguese transcript.
A Brazilian fintech translates a 17-minute Portuguese investor update into English for international shareholders.
Cost: 850 credits (85¢) with a English voice, or 17 credits (1.7¢) for English subtitles only. Portuguese is detected automatically; you only choose English.
The English voice track
English was the voice model's first language, so it is the most natural-sounding dubbing target and a good benchmark before you try other languages.
A Japanese cooking channel turns an 18-minute episode into English so international subscribers can follow along without reading subtitles.
Voice, MP4, subtitles and transcripts: 900 credits (90¢). Subtitles and transcripts only: 18 credits (1.8¢).
English subtitles
English captions are the most requested subtitle track on video platforms, and many people watch with the sound off — an English SRT or VTT makes a video usable for them.
English (English) is written in the Latin alphabet.
A German startup adds English subtitles to a 12-minute product demo before putting it on its website.
Switch Full translation output off for subtitles only: 12 credits (1.2¢). If the video is already in English, select English as the language instead and the subtitles come back untranslated — see English transcription.
Translating audio files into English
A Spanish-language podcast turns a 45-minute episode into English audio for new listeners in the US.
Podcasts and recordings go through the audio translator, which returns a English M4A and transcripts: 2,250 credits ($2.25) with voice, 45 credits (4.5¢) as text only.
Who translates into English
English is spoken in worldwide — the most common shared language for business, science and online video. It's used by creators and companies aiming at an international audience, students following lectures recorded in other languages, and researchers reviewing foreign-language interviews.
Which English ends up on screen
The language picker has one English option, with no separate US or UK setting. Spelling therefore follows whatever the machine translation produces, and a single file can drift between "color" and "colour" or "organize" and "organise". If your company has a style guide, settle on one convention and fix the SRT and VTT files to match before you publish; they are plain text, so a find-and-replace pass in any editor takes minutes.
Vocabulary drifts the same way. A German "Handy" may come out as "cell phone" or "mobile phone", and "Fußball" as "soccer" or "football". For a mainly American or mainly British audience, these are the words worth searching for. For a mixed international audience, use whichever term your website already uses so the video and the page agree.
Writing for people who learned English second
Many people watching an English version speak it as a second or third language, and that changes what good output looks like. Short sentences, one idea per line and plain verbs are easier to follow than idioms, even when the idiom is correct. Machine translation tends to keep the speaker's sentence structure, so a long, nested Italian or German sentence can arrive as a long, nested English one. In the subtitle files you can split such a line in two; the plain-language scripting guide has techniques that work just as well on translated text.
The dubbed track uses one stock voice at the pace of the original speaker. The result-screen preview plays from 0.75× to 1.5×, which is a quick way to judge whether a dense passage is followable for a non-native listener before you share the MP4.
Figures, units and dates
Machine translation usually renders what the speaker said rather than localising it. If a French engineer says "trois cents kilomètres", expect three hundred kilometres, not miles, and a Japanese price in yen stays in yen. That keeps the content accurate, but an American audience may need a conversion added to the subtitles or the video description. Dates are the other trap: "03/04" means 3 April to a European reader and March 4 to an American one, so spell months out in anything you edit by hand.
A Hamburg logistics researcher wants her keynote online in English. Dub mode with English as the target costs 1,250 credits ($1.25) and returns the English MP4, an English M4A, the German transcript, and English SRT and VTT files. If colleagues only need to read it, transcript mode with English picked costs 25 credits (2.5¢) and returns the English text plus English subtitles.
When the English transcript is the deliverable
Plenty of jobs that start as "translate this video" end as a document. Journalists checking quotes, researchers coding interviews and compliance teams reviewing supplier footage need to read and search the content, not watch it. For them, the side-by-side view of original and English text on the result screen is the working tool, and the timestamped transcript lets you find the moment a line was spoken. The article on translating foreign footage for journalism covers verifying quotes before they go to print.
Choose the dub when the audience will watch rather than read: training, product demos and explainers. The subtitles vs dubbing guide walks through that decision with more examples.
Checks before the English version goes live
- Pick one spelling convention and make the subtitles consistent with it.
- Search for names of people, products and places; names transliterated from Japanese, Korean, Arabic or Russian often have several valid English spellings.
- Read every number against the original transcript, especially prices, dates and quantities.
- Look for idioms translated word for word; they read as odd rather than wrong, so they are easy to skim past.
- Play the dub at 1× from start to finish once; listening catches problems that reading the text misses.
- If the source had several speakers, remember the dub uses one voice, so add names to the subtitles or description where it matters who is talking.
Using the tool on this page
- The spoken language is detected automatically; you select the output language — the spoken language itself for an untranslated transcript, or another language for a translation. Files can be up to 2 hours long.
- Only the audio is sent for processing — the video stays on your device — and audio and text are deleted within 30 minutes of delivery (how files are protected).
- You can download a translated video that keeps your music and effects, the translated audio as M4A, a transcript and subtitles (SRT and VTT). Every output, step and limit is explained on the video translator page.
Translating English videos into…
Other languages
Frequently asked questions
Can I translate a video to English if I don't know the original language?
Yes. The spoken language is detected automatically, so you only choose English as the target. This works for any of the other 20 languages mydubly recognises.
Do I get the Spanish transcript too?
Yes. Every translation includes the original-language transcript next to the English one.
Does it understand Hinglish?
Yes. Mixed Hindi-English speech transcribes well, and English words carry straight through to the English translation.
Will honorifics like -san be kept?
Usually not — the translation renders meaning in natural English. Add them back in the transcript if your audience expects them.
Does it work for Cantonese?
Results are less reliable than for Mandarin. Try a short clip first.
Are Korean names romanised correctly?
Usually, but romanisation varies (Lee/Yi, Park/Bak). Check names against how the people spell them in English.
Does it understand Quebec French?
Yes, Québécois speech generally transcribes well, though strong slang may be translated literally.
Does it handle Swiss German?
Standard German transcribes well. Swiss German dialect differs a lot and is less reliable.
Does it work with Egyptian or Gulf Arabic?
It can, but dialect speech is less reliable than Modern Standard Arabic. Try a short clip first.
Does it understand European Portuguese?
Yes. Both Brazilian and European Portuguese accents transcribe well.
Can I choose British or American English?
Not as a setting. There is a single English target, so spelling and vocabulary follow the machine translation. Standardise the SRT and VTT files to your house style before publishing; the dubbed audio keeps whatever wording the translation produced.
Is an English transcript good enough for a court or visa application?
Usually not. Courts, immigration offices and many contracts ask for a certified translation signed by a translator. A machine-translated English transcript is still useful for deciding which passages need that certified work; see certified translation vs AI transcripts.
What does it cost to translate a one-hour video into English?
A 60-minute dub costs 3,000 credits ($3.00). The same hour in transcript mode with English selected costs 60 credits (6¢) and gives you English text and subtitles without a voice.